Qwen3-30B-A3B-Thinking-2507-Amoral-Edition is a fusion of specialized models using the DARE TIES technique, creating a versatile solution for natural language applications.

📋 Overview

This model was developed using the DARE TIES method (Drop And REscale with Ties-Elimination), combining specialized models to create a compact and efficient solution for natural language conversations.

🔧 Base Models Used

Qwen3-30B-A3B-Thinking-2507-Amoral-Edition is the result of merging the following models:

🛠️ Merge Tool

The merge was performed using LazyMergekit, simplifying the process of merging language models with advanced configurations.

🧩 Technical Configuration

Merge Parameters

yaml

models:
  - model: Ewere/Qwen3-30B-A3B-abliterated-erotic
    parameters:
      density: 0.6
      weight: 0.6
  - model: unsloth/Qwen3-30B-A3B-Thinking-2507
    parameters:
      density: 0.6
      weight: 0.4
merge_method: dare_ties
base_model: unsloth/Qwen3-30B-A3B-Thinking-2507
parameters:
  normalize: true
  int8_mask: false
dtype: bfloat16

Technical Specs

Architecture: Qwen3 30B
Merge Method: DARE TIES
Precision: BFloat16
Normalization: Enabled
Int8 Mask: Disabled
Language: English

💻 How to Use

Dependency Installation

pip install -qU transformers accelerate torch

Basic Example

python

from transformers import AutoTokenizer
import transformers
import torch

model = "rodrigomt/Qwen3-30B-A3B-Thinking-2507-Amoral-Edition"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)

pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])

English Conversation Example

python

conversation = [
    {"role": "user", "content": "Hello! How are you?"},
    {"role": "assistant", "content": "Hi! I’m doing well, thanks for asking. How can I help you today?"}
]
prompt = tokenizer.apply_chat_template(conversation, tokenize=False, add_generation_prompt=True)
outputs = pipeline(prompt, max_new_tokens=128, temperature=0.7)
print(outputs[0]["generated_text"])

⚠️ Minimum Requirements

RAM: 16GB
VRAM: 24GB+ (GPUs with less will not run this model properly)
Storage: 20GB available
GPU: RTX 3090 / A6000 / H100 or higher